CircleID

I have not seen a more detailed and rational description of the issues involved here. Great job, Geoff.

Excellent. Though the origins of most of the opposition to the current structure can be summarized like this, is what I suspect:

* Large government owned telco + ISP is not too good at managing current address space, and finds it difficult to allocate new address space

* Customers of that ISP find it so difficult to prise addresses out of that ISP without high costs in bureaucracy and time that they resort to NAT'ting large networks, trying to run research labs off a single IP where they'd have at least a /24 elsewhere.

* The myth of IP address shortage spreads and is laid at the RIR's door

* The government owned telcos provide qualified experts in telecom public policy to their governments, which then recommend something like this.

There you go .. recipe for disaster

Great introduction.

It would seem that the critical issue that Zhao's proposal intends to address is one of *allegedly* poorly run RIRs. 

It would seem that the critical issue of the proposal is the consequences of the incentive for policy dilution that it creates.

I think most observers and fair stakeholders will see that this proposal would be disastrous, but can its motivations be better addressed?

Are the RIRs policies set up to make it as clear as possible that they are being responsive?  It's rather difficult to make it clear that the RIRs, whose jobs are to sometimes say no to IP space requests, are nevertheless doing their jobs well, despite the dissatifaction that these 'NOs' often produce.  Would greater openness *in fact* help without compromising their task?  Can we contrast, e.g. RIR, ICANN, and IETF policies and their results?

The irony is that the RIRs are well run, for the most part. ICANN, the ITU and other more normative International organizations are, on the other hand, not in the same ball park, especially in terms of effectiveness.

Where is the line between deliberate falsehood and passionate advocacy? The "debate" surrounding RIRs and IPv6 is chock full of demonstrably untrue supositions such as the allegation of unfair address distribution. Do the folks who hold to such ideas realize that they are factually incorrect or are they well meaning but misinformed?

Suresh's comments concerning the PTTs are right on the mark - this is the result of government sponsored telecommunications entities that see addressing as another element to add to their quasi-monopolies. The RIRs present too uniform of a playing field, which clearly disagrees with government sponsored telco's, for whom self-identity and nationalism are uncomfortably intertwined. 

Even if Zhao's proposal is not the best technique for the assignment of IP addresses, it would be hightly desirable to bind each Internet user of an IP address to a National Governament.

Currently, we grab the country that an IP belongs to using techniques that are not official, not standardized, and not reliable. That is not sufficient for legal bindings, nor for applying any policy that requires legal support. We need an efficient and reliable way of determining an IP's Country of origin. Then, it is a technical point to reckon if assigning IPs on a per-Country basis is better or worse than an enhanced reverse lookup.

Let's not confuse those two points of view.

In case it wasn't clear: I agree that the RIRs are both relatively and absolutely well-run.  I think it's appropriate to assume the line has been crossed into deliberate falsehood, and then figure out how to go from there.  Recent fabricated stories scandals suggest the mainstream media hasn't figured this out.

Ale: Given the expected continued prevalence of anonymous proxies, I see a (hypothetical) zero-cost, perfect IP:nation-state mapping as being only very slightly more useful than the current system; it would not provide the legal certainty you desire.

I can forsee Zhao's scheme leading to repetitions of past situations, (which have occurred in many if not most nations-the US and Peru come to mind, and there are probably contemporary examples) where getting a new # took months or years.

In the words of the Draft Declaration of Principles (verset 49/a) "policy authority for Internet-related public policy issues is the sovereign right of States. They have rights and responsibilities for international Internet-related public policy issues."

Thus the problem is not much with proxies, as they provide for explicit directives to get in direct touch with the user. Rather, the point is that National Governaments should take responsibility and liability for what their citizens do with IP addresses. It is a political question, but it requires clear and official statements as to which numbers have been assigned to which Nation. IMHO, that can be accomplished with an enhanced reverse-DNS or similar technology.

Zhao's proposal does not solve the problem of mapping IP addresses to Nations, since it provides for only a part of the IPv6 address space being assigned directly to National units, thereby missing IPv4 and the rest of IPv6.